METHOD FOR LASER PROCESSING OF METALLIC MATERIALS BASED ON AUTOMATIC DETERMINATION

Abstract

A machine and associated method for laser processing of metallic materials includes: (a) controlling emission of at least one pulse of a characterization laser beam onto a predetermined region of the material in a characterization atmosphere to generate metal vapor and / or plasma from the material; (b) acquiring spectral data representing an emission spectrum of the metal vapor or plasma indicative of the material being processed; (c) identifying, by electronic processing and automatic recognition means configured in a supervised learning phase with a set of training spectral data samples indicative of the predetermined material classes or predetermined material processing parameter classes, one of a plurality of predetermined material classes or predetermined processing parameter classes corresponding to the acquired spectral data; and (d) selecting current processing parameters for the material in accordance with the identified material class or processing parameter class.

Description

[Technical Field]

[0001] The present invention relates to the laser processing of materials, more particularly metallic materials, and in particular to improvements to methods and machines for laser processing of materials for laser cutting, drilling or welding of materials or for additive manufacturing of three-dimensional structures from powders of materials.

[0002] More particularly, the invention relates to a method for laser processing of metallic materials according to the preamble of claim 1.

[0003] According to a further aspect, the invention relates to a machine for laser processing of metallic materials according to the preamble of claim 18. [Background technology]

[0004] In the following specification and claims, the term "metallic material" is used to identify any product, such as a sheet or elongated profile, having a closed cross section (e.g., hollow circular, rectangular or square shape) or an open cross section (e.g., flat cross section, or L-shaped, C-shaped or U-shaped, H-shaped, I-shaped cross section, etc.).

[0005] In the industrial processing of materials, especially metal sheets and profiles, lasers are used as heating tools for a wide range of applications, which depend on the interaction parameters of the laser beam with the material to be processed, in particular the incident volumetric energy density of the laser beam on the material and the interaction time interval.

[0006] Laser processing machines are widely known and include industrial processing machines equipped with CO2 or ytterbium-doped fiber laser sources or direct diode laser sources. These sources are configured to emit single-mode or multimode laser beams, with the optical path of the laser beam being in air or in an optical fiber. The laser beam is guided from the source along a beam optical transport path to a processing head. The processing head includes optics for focusing the laser beam along its optical propagation axis for incidence on the material.

[0007] Low energy density (surface mm 2 The hardening process is carried out by directing a laser beam with a high energy density (of the order of tens of watts per mm) at the metal material for a long time (of the order of a few seconds), while 2 Ablation processes are performed by directing a laser beam with a power (on the order of tens of megawatts per second) at the same metallic material for times on the order of femtoseconds or picoseconds. In the intermediate range of increasing energy density and decreasing processing time, control of these parameters allows welding, cutting, drilling, etching, and marking processes to be performed.

[0008] The differences between the various types of processing that can be performed on materials therefore depend substantially on the power of the laser beam used and the interaction time between the laser beam and the material to be processed.

[0009] Also, in additive manufacturing, control of the power of the laser beam and the interaction time of the laser beam with the material is crucial to achieve the processing goals and ensure the desired properties of shape, compactness and stability of the resulting structure.

[0010] Therefore, within the framework of a predetermined type of processing, the quality of the processing depends on the setting of the interaction parameters between the laser beam and the material to be processed, which obviously depends on the type of material to be processed and the conditions at the time of processing (e.g. surface, environmental conditions, etc.).

[0011] In fact, different metal alloys react differently to the processing laser beam and encounter further differences in the execution of the processing as a result of the surface state of the material (e.g., oxidation state, the presence of metal surface coating layers or other surface modifications such as protective oil layers) and the processing atmosphere.

[0012] However, it is often not possible to know the actual composition of a metal alloy precisely, and it may change over time and may vary relative to that indicated by the manufacturer as a result of storage conditions. Furthermore, the processing of materials may use various types of assist gases with different parameters depending on the state of the material.

[0013] There is therefore a desire to improve the quality of industrial processes of laser machining by adapting the machining parameters as efficiently and quickly as possible to the material to be machined and the actual machining conditions. Summary of the Invention [Problem to be solved by the invention]

[0014] An object of the present invention is to provide a method for laser processing of metal materials that is optimized for the material and processing conditions.

[0015] It is a further object of the present invention to provide an optimized method for laser processing of metallic materials that can be efficiently implemented in a production line. [Means for solving the problem]

[0016] According to the invention, these objects are achieved by a method for laser processing of metallic materials having the features of claim 1.

[0017] Particular embodiments form the subject matter of dependent claims, the content of which is to be understood as an integral part of this description.

[0018] A further subject of the invention is a machine for laser processing of metallic materials as defined in the claims.

[0019] In summary, the present invention is inspired by the observation that the type of material (typically a metal alloy) undergoing processing can be qualitatively estimated from the phenomenon of metal vapor or plasma (or a combination thereof) emitted by a material being or already being processed (detected by acquiring an emission spectrum in a spatial region around and near the material). The term "material being or already being processed" refers to the material exposed to the processing laser beam, as well as the material's surface layer (due to a finish, coating, or unwanted oxidation) or subsurface layer exposed during processing. The term "emission" refers to detectable light emission across a broad electromagnetic spectrum, including the ultraviolet and infrared regions, resulting from ionization of the metal vapor or plasma, or, in the case of infrared, indicating thermal radiation. The collected spectra are used for the identification of the material being processed, not for process monitoring or two-dimensional or three-dimensional mapping.

[0020] In particular, the invention is inspired by a technique for the spectroscopic characterization of materials known as LIBS (Laser-Induced Breakdown Spectroscopy), which allows the identification of the chemical composition of a sample, either qualitatively or quantitatively (the relative proportions of different elements in the sample) at the atomic level, by analysis of the light spectrum obtained by using a laser to generate a plasma from the material.

[0021] The operation of this technique is diagrammed in Figure 1. A short laser pulse L is focused on the sample (material) M, and the interaction of the radiation with the material forms a plasma plume P. Upon recombination of ions and electrons, radiation R is emitted in the optical band from ultraviolet to near-infrared. Part of this is continuum radiation, while the other part is characteristic of the chemical elements that make up the sample. The radiation is then collected via an optical line and guided to a spectrometer S, where it is split into different wavelengths. The result is an emission spectrum U that is specific to the material, but is influenced by the laser used and the atmosphere in which the process occurs. The collected spectrum is analyzed and processed using appropriate algorithms to quantitatively trace the chemical composition of the material.

[0022] Lasers used for material characterization by LIBS are typically pulsed on the order of 10 nanoseconds. 10 ~10 15 W / cm 2 Lasers with femtosecond-order pulses can be used in vacuum laboratories, but the most common portable systems are based on Nd:YAG lasers with emission at a wavelength of 1064 nm in air, which only in some cases have the ability to operate in inert atmospheres (such as argon).

[0023] The collected spectra are analyzed and processed by comparison with existing data sets or standard samples for precise calibration of the instrument, thereby providing a quantitative assessment of the chemical composition of the material.

[0024] In an innovative manner, the method of the present invention uses LIBS techniques to control laser processing of materials. Specifically, the laser processing of materials is controlled by selecting processing parameters for the material according to the class of processing parameters associated with characteristic spectral data of the emission of metal vapor or plasma (or a combination thereof) generated by the material, i.e., identified by automatic recognition from such spectral data. The automatic recognition is based on supervised learning of correspondences between a set of training spectral data samples and the associated material classes or processing parameter classes.

[0025] The method covered by the present invention can be implemented by a system adaptable to a laser processing machine, which can be easily integrated therein and share part of the physical hardware (i.e. the laser source and part of the optical path adapted to perform the processing) or processing resources already present in the machine and provided for controlling the laser beam and the entire laser processing. Advantageously, the method covered by the present invention can be implemented by using the laser beam generated by the laser source in an innovative way as a processing tool to generate metal vapor or plasma emissions from the material, which are necessary for its characterization.

[0026] Advantageously, the quality of laser processing is improved as a result of the in-line (i.e., not in a vacuum, but possibly in a protective atmosphere) identification of optimal processing parameters that most closely approximate (or match) the machining parameters employed in a previous process applied to a material with the same (or closest) spectroscopic signature, i.e., the same (or closest) type of material to be processed. In-line identification of processing parameters is advantageously carried out immediately before processing, either directly or via material identification, and allows an automatic and more accurate calibration of the parameters required for processing, allowing any errors in the material pre-determination to be corrected and the calibration to be adapted to the bulk material or to a specific surface treatment.

[0027] Even more advantageously, the techniques covered by the present invention make it possible to detect the metal material to be processed on the machine during laser processing of the material and continuously during the process, without the need for downtime or prior data acquisition (especially visual acquisition by the operator) of the material received from the supplier.

[0028] Finally, the invention makes it possible to improve the laser processing process of metal materials as a result of material recognition directly in the machine, making it possible to identify errors in the pre-specification of the material to be processed, avoiding production waste, reducing production costs and optimizing the expected quality of the finished product. [Brief explanation of the drawings]

[0029] Further features and advantages of the invention will be set out in more detail in the following detailed description of embodiments of the invention given as non-limiting examples, with reference to the accompanying drawings, in which: FIG. [Figure 1] The operation diagram of the LIBS method is shown below. [Figure 2a] Three different configurations of machines for laser processing of metallic materials operating in accordance with the method covered by the present invention are shown schematically, with identical or functionally equivalent elements or components being designated by the same reference numerals. [Figure 2b] Three different configurations of machines for laser processing of metallic materials operating in accordance with the method covered by the present invention are shown schematically, with identical or functionally equivalent elements or components being designated by the same reference numerals. [Figure 2c] Three different configurations of machines for laser processing of metallic materials operating in accordance with the method covered by the present invention are shown schematically, with identical or functionally equivalent elements or components being designated by the same reference numerals. [Figure 3] 1 is a flow chart of operations performed to prepare a metal material for laser processing in accordance with methods covered by the present invention. [Figure 4a] 1 is an optical plasma emission spectrum of an exemplary material. [Figure 4b] 1 is an optical plasma emission spectrum of an exemplary material. [Figure 4c] 1 is an optical plasma emission spectrum of an exemplary material. [Figure 4d] 1 is an optical plasma emission spectrum of an exemplary material. [Figure 4e] 1 is an optical plasma emission spectrum of an exemplary material. [Figure 5] FIG. 1 is a schematic diagram of a preferred classification method employed in methods covered by the present invention. [Figure 6] Projection of spectral data in a two-dimensional classification space defined by two principal components, representing four material classes. [Figure 7] We present the results of the analysis of the classification space in Figure 5, which maximizes the margin between classes. [Figure 8] A new spectral data class is shown in the analysis results of Figure 6 in the state of characteristic anomalies. DETAILED DESCRIPTION OF THE INVENTION

[0030] FIG. 1 has been described above, and its contents are intended to be referenced herein as a general guide to the configuration of a controlled processing machine for carrying out the method according to the teachings of the present invention.

[0031] FIG. 2a is a schematic diagram of a machine for laser processing of metallic materials according to the invention in a first configuration.

[0032] The machine of Fig. 2a comprises at least one laser emission source 10 configured to emit a laser beam B into a conveying means, for example an optical fiber 12, which is configured to guide the processing laser beam emitted from the source towards a processing head 14 located in the vicinity of the material WP. The processing laser beam B is, for example, a continuous wave laser beam having a peak power of 1 kW or more.

[0033] A laser emission control unit 16 can be connected to the light source 10, configured to control the power and emission duration parameters of the laser beam B and / or select one of the multiple laser beams in the case of multiple light sources. In one embodiment, the control unit 16 is configured to control the emission of at least one pulse of the characterization laser beam onto a predetermined area of ​​the material WP, which is configured to generate metal vapor and / or plasma, with subsequent optical and thermal emissions representing a "signature" of the material. As a result of the power and / or pulse duration of the characterization laser beam, the material undergoes a gradual heating process, which results in partial evaporation and subsequent possible excitation and ionization to form plasma. In the space surrounding the point of incidence of the characterization laser beam on the material, an atmosphere is formed in which a combination of metal vapor and plasma (known as a "plume") or either alone can exist. The characterization laser beam may originate from a light source (not shown) separate from light source 10 and be guided separately to the processing head, in which case the characterization laser beam is conveniently guided along a common optical path with the processing laser beam.

[0034] In a currently preferred embodiment, a single light source 10 is used, and the characterization laser beam is obtained by modulating the continuously emitted processing laser beam, for example by modulating the processing laser beam to emit one or more laser pulses with durations of 10 to 200 microseconds, thus eliminating the additional cost of a second light source in the machine.

[0035] The processing head 14 typically includes an optical system with a set of optical elements such as, for example, a collimating lens 20, a beam-shaping mirror 22, a focusing lens 24, and protective optics 26.

[0036] The optical combiner 28 is attached to the light source and configured to combine and launch the multiple laser beams onto a transmission fiber and to direct optical signals emitted from the metal vapor or plasma of the material and collected by the processing head optics to a spectral data acquisition means 30, such as a spectrometer, configured to acquire spectral data representing the emission spectrum of the metal vapor or plasma generated by the characterization laser beam. These spectral data are indicative of the material WP being processed. Although not shown, the optical signal path downstream of the combiner 28 toward the spectrometer 30 may include an optical collimation and filtering system.

[0037] The processing head 14 is fitted with optical actuator means, illustrated in the figure as actuator means 32 for moving the beam-shaping mirror 22 and actuator means 34 for beam focus adjustment acting on the collecting lens 24, both of which are controlled by a control unit, generally designated 36, and connected to a field bus 38, to which is also connected processing and control means 40 configured to control the irradiation of the processing laser beam along a predetermined processing path. Controlling the irradiation of the processing laser beam along a predetermined processing path includes controlling the supply of an assist gas flow (in cutting applications) and controlling the irradiation of a predetermined power distribution of the laser beam towards a predetermined processing area by reference to a predetermined processing model or program, i.e. according to processing trajectory information and processing parameters obtained in the form of movement instructions for the processing head and / or the material to be processed, as well as physical processing parameters indicating the optical beam power distribution, the beam power intensity and the activation time of the laser beam as a function of the processing trajectory. The processing means 40 may be integrated into a single processing unit mounted on the machine or may be implemented in a distributed manner, in which case they comprise a number of processing modules installed in different parts of the machine, including, for example, the processing head.

[0038] The main advantages of this configuration are the non-intrusive installation of the spectroscopic data acquisition means, the elimination of the need for modification or custom design of the processing head to accommodate the spectrometer, and the overall system being more robust due to the absence of sensitive metrology equipment near the processing area. These considerations make this configuration more preferable in industrial situations. Because the optical fiber 12 used as the transmission means does not transmit light in the ultraviolet range, emission at wavelengths below 400 nm may not generally be observed in this configuration. Furthermore, emission peaks will be present at the emission wavelengths of the characterization and processing laser beams.

[0039] In an alternative embodiment to that described above, including an optical coupling device 28, shown in FIG. 2b, in which elements or components identical or functionally equivalent to those shown in FIG. 2a are designated by the same reference numerals, the spectral data acquisition means 30 is separate from the laser light source and is integrated into the processing head 14.

[0040] In this configuration, optical signals emitted from the metal vapor or plasma of the material are collected by the processing head optics, including protective optics 26 and collecting lens 24, and transmitted through dichroic beam shaping mirror 22 towards spectral data acquisition means 30 integrated into processing head 14. The path of the optical signal downstream of dichroic mirror 22 to spectrometer 30 may include collimating and filtering optics 30'.

[0041] In both cases, the collection line that allows the acquisition of the optical signal emitted by the metal vapor or plasma includes a spectrometer that uses a very wide band (ultraviolet to near infrared) or only a part of it. Alternatively, sensor means, such as multiple photodiodes, can be used that are intended to determine the signal intensity only at predetermined wavelengths in the near infrared to ultraviolet optical band.

[0042] Advantageously, in the two configurations described, the collection of the optical signal into the spectral data acquisition means 20 is performed according to a coaxial configuration, with the optical signal following the same optical propagation path as the characterization and processing laser beam for at least a certain distance, passing through the processing head and, more advantageously, reaching an optical coupling device 18 attached to the light source. Using the same optics to focus the characterization (and processing) laser beam and to acquire the plasma emission ensures that the emission is collected from the same area of ​​the material where the plasma was generated.

[0043] Alternatively, as shown in Figure 2c, where identical or functionally equivalent elements or components to those shown in Figure 2a are designated by the same reference numerals, depending on the space available in the machine, optical signal collection can be performed off-axis outside the processing head via lateral positioning of the spectral data acquisition means 30 at a distance and angle relative to the predetermined propagation direction of the characterization and processing laser beam. This configuration allows for the collection of spectral data that is not distorted or filtered by the optical shaping system of the characterization and processing laser, thereby preserving the natural shape of the emission and allowing spectral data to be acquired even in the ultraviolet range. In either case, it is preferable to place a cutoff filter for the reflected wavelength of the laser beam on the material to avoid saturating the emission spectrum. Unfortunately, because metal vapor or plasma emission from the material is not collected via the optical collection system of the characterization (and processing) laser beam, changes in the processing head position relative to the material (e.g., changes in processing distance from the material) require adjustments to the distance and positioning angle of the spectral data acquisition means relative to the propagation direction of the characterization and processing laser beam.

[0044] In each of the above configurations, an electronic processing and automatic recognition means 42 connected to a data logging memory DB is connected to the field bus 38 and configured to process the signals emitted from the spectral data acquisition means 30 and identify one of a plurality of predetermined material classes or predetermined processing parameter classes corresponding to the spectral data. The electronic processing and automatic recognition means 42 is configured for a supervised learning phase using a training spectral data sample set representing the predetermined material classes or predetermined material processing parameter classes. The training spectral data sample set may be obtained based on previously processed and periodically updated material nameplate data, or may be obtained through preliminary acquisition of spectral data samples, such as several dozen spectra, for each available material. The electronic processing and automatic recognition means 42 is configured to select processing parameters corresponding to the processing parameter classes or material classes identified by the data logging memory DB, which includes a reference model showing the nominal relationship between the spectral data and the processing parameter classes or material classes.

[0045] Generally, the term "processing parameters" refers to at least one of the physical parameters of the processing laser beam and the processing environment, such as the repetition rate of the processing laser beam pulses, the duration of the processing laser beam pulses, the power of the processing laser beam pulses, the power density distribution of the processing laser beam, and the pressure of the assist gas, or more generally to a processing strategy that includes one or more of the physical parameters of the processing laser beam and the processing environment, and control parameters of the movement of the processing beam.

[0046] The processing control means 40 is configured to control the irradiation of the processing laser beam along a predetermined processing trajectory in accordance with selected processing parameters.

[0047] A simplified flow chart of the method covered by the present invention is shown in FIG.

[0048] In step 100, the control unit 16 is initiated by the processing and control means 40 or the electronic processing and automatic recognition means 42 to control the emission of at least one pulse of the characterization laser beam onto a predetermined area of ​​the material WP in the characterization atmosphere to generate metal vapor and / or plasma from the material. In certain cases, the characterization atmosphere is the same as the atmosphere of the processing process.

[0049] The control unit 16 controls the emission of at least one pulse of the characterization laser beam during an initial stage of processing the material, for example, onto an area of ​​the material to be cut or drilled, or onto an area of ​​the material not to be processed during a calibration step before processing the material.

[0050] The control unit 16 may control the emission of at least one pulse of the characterization laser beam during processing of the material, conveniently in the case where the characterization laser beam is obtained by modulating the processing laser beam, possibly by changing the processing atmosphere to temporarily provide a different characterization atmosphere.

[0051] In step 200, the spectral data acquisition means 30 acquires spectral data representing the emission spectrum of a metal vapor or plasma indicative of the material WP to be processed.

[0052] As an example, the spectra of aluminum alloy Al6060 (Figure 4a), galvanized steel (Figure 4b), and stainless steel AISI304 (Figure 4c) are shown in Figures 4a-4c. Various intensity values ​​of the emission can be detected depending on the plasma generation conditions of the material (e.g., due to the power of the characterization laser beam, the process atmosphere, and more generally, the laser processing machine used). Figure 4d shows the spectra of aluminum alloy Al6060 in an oxygen atmosphere obtained from characterization laser beams emitted from various laser processing machines with various powers and diameters. Various spectra can be obtained depending on the surface condition of the material (e.g., oxidation state and its coating). Figure 4e shows a comparison between the spectrum of carbon steel S235JR in normal condition (curve A) and the spectrum of rusted carbon steel S235JR (curve B).

[0053] In step 300, the electronic processing and automatic recognition means 42 is pre-configured in a supervised learning step not shown to identify either a plurality of predetermined material classes or predetermined processing parameter classes corresponding to the acquired spectral data.

[0054] Before applying the automatic recognition algorithm to the spectral data, it is useful to normalize the data in order to have comparable data, since even in the presence of the same parameters of the characterization laser beam, there may be some variation in the overall intensity of the plasma emission. Due to the spectral normalization, it is possible to compensate for variations in the intensity of the acquired spectral data.

[0055] Each spectrum is a set of N emission lines I(λ) acquired in a predetermined wavelength range between ultraviolet and infrared (e.g., 181 nm to 1100 nm, average resolution 0.56 nm). i ) (i=1,…,N).

[0056] Conveniently, two normalization procedures can be employed: First, the spectrum is normalized by dividing all emission line intensities by the maximum intensity (in other words, the maximum peak is represented by a value of 1 and the other values ​​are scaled accordingly) according to the following formula:

number

[0057] The second step is to divide each intensity of the emission line by the sum of all intensities (essentially scaling the individual values ​​relative to the total emission represented by the area enclosed by the spectral curve) according to the formula:

number

[0058] In step 400, the electronic processing and automatic recognition means 42 selects the current processing parameters of the material according to the material class or the recognized processing parameter class.

[0059] Finally, in step 500, the processing and control means 40 controls the application of the processing laser beam at a predetermined area of ​​the metal material or along a predetermined processing trajectory according to the current processing parameters selected in the previous step.

[0060] Conveniently, to enable recognition of materials over large surface areas or depth volumes where material conditions may vary, step 100 may be repeated continuously or at predetermined processing intervals, whereby control unit 16 controls the application of multiple pulses of the characterization laser beam to a series of predetermined regions of a two-dimensional surface scan area of ​​the material or a series of predetermined regions of a three-dimensional scan volume of the material to generate individual metal vapors or plasmas from the material present in such scan areas. For example, control unit 16 may control the application of multiple pulses of the characterization laser beam to a series of depths of a predetermined region of a three-dimensional scan volume of the material, and focus the characterization laser beam on different surfaces of the material to include a surface layer of the material and at least one sub-surface layer exposed as a result of characterization of the upper layer.

[0061] More particularly, the electronic processing and automatic recognition means 44 may be suitably implemented as a recognition and classification module based on principal component analysis (PCA) techniques for feature extraction and artificial intelligence classification systems such as neural networks or support vector machines (SVM). The spectral data may be processed through various statistical methods, such as PCA, as well as through normalization techniques, for example by using spectral data obtained across the entire detected spectrum.

[0062] A schematic diagram of the classification method described above is shown in Figure 5.

[0063] In a currently preferred embodiment, identification of a material class or a class of processing parameters corresponding to the acquired spectral data is performed by transforming the acquired spectral data into a classification space defined by predetermined orthogonal latent variables comprising a predetermined subset of significant latent variables indicative of the variance of the spectral data, and comparing n sets of values ​​of the significant latent variables calculated from the detected spectral data with a set of reference values ​​of the n sets of significant latent variables indicative of a set of training spectral data samples. Parameters corresponding to the predetermined processing parameters associated with the reference values ​​of the n sets of significant latent variables that have a predetermined metric relationship with the n sets of values ​​of the significant latent variables calculated from the detected spectral data are selected as processing parameters for the material.

[0064] Conveniently, the predetermined metric relationship is a minimum distance relationship.

[0065] Figure 6 shows the projection of spectral data acquired for a particular characterization laser beam and a particular characterization atmosphere in a two-dimensional classification space defined by principal components identified as PC2 and PC3. The training spectral data samples contain four classes: Al6060 (aluminum, magnesium, and silicon-based alloy), AISI304 (stainless steel), E220 (steel) characterized in an O2 atmosphere, and E220 (steel) characterized in an N2 atmosphere. As can be seen graphically, the space of the two principal components, PC2 and PC3, clearly distinguishes the four classes of the training spectral data samples.

[0066] Figure 7 shows the spatial partitioning of the principal components PC2 and PC3 resulting from the application of the SVM classifier. This partitioning results from using a linear kernel and a partitioning function that maximizes the margin between the classes.

[0067] As with any type of classifier, each new measurement, even if it is a new material or new experimental condition, is classified into one of the known classes (those used to train the model). As is evident from Figure 7, there are regions far removed from the clusters observed in training, where new observations are forced to be classified into one of the known classes despite significant deviations from the training data. Therefore, improvements to this method can be introduced by extending the classification and reporting whether a given classification result is reliable. This can be done, for example, by automatically determining whether the spatial projection of the principal components of a new observation is within (or near) the cluster of training data belonging to the assigned class, or deviates from it by a distance greater than a predetermined threshold, and reporting this characterization anomaly to the operator. Figure 8 shows an example in which a new observation class N is characterized by a score that lies within the spatial region where the material is classified as an Al6060 system, but the acquired spectral data should be determined to deviate significantly from the corresponding cluster. A characterization anomaly condition is confirmed when the distance of each representation of acquired spectral data, or the distance of the center of a representation of spectral data acquired in multiple observations, from the center of a training spectral data cluster (or clusters of all known training spectral data classes) is greater than a predetermined threshold.The principle of distance calculation from a reference cluster in a multidimensional space is based on the same mechanism underlying the SVM method, in its variant known as one-class SVM or support vector data description (SVDD), as explained in (Tax, DM, Duin, RP, in “Support vector data description. Machine learning”, 54(1), pp. 45-66, 2004, by Ning, X., Tsung, F., “Improved design of kernel distance-based charts using support vector methods”, IIE transactions, 45(4), pp. 464-476, 2013 and from Grasso M., Colosimo BM, Semeraro Q., Pacella M., “A Comparative Study of Distribution-Free Multivariate SPC Methods for Multimode Data”, Quality & Reliability Engineering International, 31(1), pp. 75-96, 2015).

[0068] Characterization anomaly conditions can be tracked by investigating with the operator whether it is a new material or a new combination of material and processing atmosphere. If so, such data can be added to the reference model as new training spectral data samples, and the classifier can be retrained once a sufficient number of observations belonging to that class are available. If not, it may still be useful to include new observations in the training sample set for that material to account for additional variation not observed in the initial training phase. This allows for improved classification performance in the future. Also, in this case, if the training spectral data sample set is expanded, the classification algorithm will need to be retrained, and the distance threshold value that defines the conditions for reporting a characterization anomaly will need to be redefined.

[0069] Naturally, without prejudice to the principles of the invention, the embodiments and details of implementation can be widely modified from those described and illustrated purely by way of non-limiting example, without departing from the scope of protection of the invention as defined by the appended claims.

Claims

1. 1. A method for laser processing of metallic materials, in particular for laser cutting, drilling or welding volumes of said material or for additive manufacturing of three-dimensional structures from powders of said material, comprising: irradiating a processing laser beam at a predetermined area or along a predetermined trajectory to process a metal material in a predetermined processing atmosphere according to current processing parameters selected based on the material; - controlling the emission of at least one pulse of a characterization laser beam onto a predetermined area of ​​the material in the characterization atmosphere to generate metal vapor and / or plasma from the material; - acquiring spectral data representing an emission spectrum of a metal vapor or plasma indicative of a material processed in said characterization atmosphere; using a set of training spectral data samples representative of predetermined material classes or predetermined material processing parameter classes, identifying, by electronic processing and automatic recognition means configured in a supervised learning phase, one of a plurality of predetermined material classes or predetermined processing parameter classes corresponding to said spectral data; - selecting current processing parameters for the material depending on the identified material class or processing parameter class.

2. 2. The method of claim 1, wherein the predetermined processing parameters include at least one of a pulse repetition frequency of the processing laser beam, a pulse duration of the processing laser beam, a pulse power of the processing laser beam, a power density distribution of the processing laser beam, and an assist gas pressure.

3. 3. The method of claim 1, comprising the step of controlling the emission of at least one pulse of the characterization laser beam at an area of ​​the material to be cut or drilled during an initial stage of processing the material.

4. 3. The method according to claim 1, further comprising the step of controlling the emission of at least one pulse of the characterization laser beam to an area of ​​the material that is not to be processed in a calibration step before processing the material.

5. 5. A method according to claim 3 or 4, comprising controlling the emission of a plurality of pulses of a characterization laser beam at a series of predetermined regions of a two-dimensional scanned volume of a material, or at a series of predetermined regions of a three-dimensional scanned volume of a material, to generate individual metal vapors or plasmas from material present within the scanned regions.

6. controlling the emission of a plurality of pulses of the characterization laser beam at a series of predetermined regions of the three-dimensional scanned volume of the material includes focusing the laser beam at different surfaces of the material, including a surface layer and at least one sub-surface layer of the material; 6. The method of claim 5, wherein a first pulse generates a metal vapor or plasma from a material in a surface layer and at least a second pulse generates a metal vapor or plasma from a material in a sub-surface layer.

7. The method according to any one of claims 1 to 6, wherein the spectral data representing the emission spectrum of the metal vapor or plasma of the material is acquired in the optical band between the near infrared and the ultraviolet.

8. 8. The method of claim 7, wherein the spectral data representing the emission spectrum of the metal vapor or plasma of the material is acquired at a predetermined wavelength within the optical band between near infrared and ultraviolet.

9. The method according to any one of claims 1 to 8, wherein the characterization atmosphere is a processing atmosphere.

10. The method according to any of the preceding claims, wherein the characterization laser beam is obtained by modulation of a processing laser beam.

11. 11. The method of claim 10, wherein the at least one pulse of the characterization laser beam is emitted for a duration of 10 to 200 microseconds by modulation of a continuously emitted processing laser beam with a peak power of 1 kW or more.

12. The method according to any of the preceding claims, wherein the characterization laser beam is guided along a common optical path with a processing laser beam in a processing head of a machine for implementing a laser processing process.

13. 12. A method according to claim 10 or 11, comprising detecting an emission spectrum of a metal vapor or plasma coaxial with the propagation direction of the processing laser beam in a processing head of a machine for implementing the laser processing process.

14. 12. The method according to claim 10 or 11, comprising detecting the emission spectrum of the metal vapor or plasma outside the processing head at a predetermined angle relative to the propagation direction of the processing laser beam in the processing head of a machine for implementing the laser processing process.

15. A method according to any preceding claim, comprising controlling the emission of at least one pulse of the characterization laser beam during processing of the material.

16. identifying, by electronic processing and automatic recognition means, one of a plurality of predetermined material classes or processing parameter classes corresponding to said spectral data, transforming the spectral data into a classification space defined by predetermined orthogonal latent variables, the latent variables comprising a subset of predetermined significant latent variables that indicate the variance of the spectral data; 16. The method of any preceding claim, comprising: comparing the n-set of values ​​of the significant latent variables calculated from the spectral data with a set of reference values ​​of the n-set of significant latent variables indicative of the set of training spectral data samples, whereby material processing parameters are selected that correspond to predetermined processing parameters associated with the n-set of reference values ​​of latent variables that have a predetermined metric relationship with the n-set of significant latent variable values ​​calculated from the collected spectral data.

17. The method of claim 16 , wherein the predetermined metric relationship is a minimum distance relationship.

18. A machine for laser processing of metallic materials, in particular for laser cutting, drilling or welding volumes of said material or for additive manufacturing of three-dimensional structures from powders of said material, comprising: a light source (10) emitting a processing laser beam (B); means for guiding the processing laser beam (B) emitted by said light source (10) along an optical path for conveying the beam (B) to a processing head (14) placed in the vicinity of said workpiece (WP); Processing and control means (40) configured to control the irradiation of said processing laser beam (B) along a predetermined processing trajectory on a material (WP) in a predetermined processing atmosphere and according to current processing parameters selected based on said material; a light source (10) for emitting at least one pulse of a characterization laser beam onto a predetermined area of ​​the material in the characterization atmosphere so as to generate a metal vapor and / or plasma from the material; - means (30) for acquiring spectral data representing an emission spectrum of a metal vapor or plasma indicative of a material processed in said characterization atmosphere; an electronic processing and automatic recognition means (42) configured in a supervised learning phase with a set of training spectral data samples indicative of predetermined material classes or predetermined material processing parameter classes, and configured to identify one of a plurality of predetermined material classes or predetermined processing parameter classes corresponding to the acquired spectral data; Thereby, the processing and control means (40) is configured to control the irradiation of the processing laser beam (B) according to current processing parameters selected according to the identified material class or processing parameter class.

Images